YouTube AI disclosure: which avatar pipeline steps are exempt?
YouTube exempts scripts, captions and upscaling from its AI label but not realistic synthetic people. Map each step of a Sume avatar pipeline to the rule.
Short answer
Most of an avatar video pipeline falls under production help that YouTube does not ask you to disclose: writing the script, generating captions, and enhancing quality. The one step that can need a label is the realistic synthetic person on screen. This post walks a Sume Avatar 1.0 pipeline step by step against the YouTube help page, read on 2026-10-08.
Step by step
The table separates each pipeline step from the part of the YouTube page it touches. The page names "script generation, thumbnail creation, caption generation" and enhancement such as sharpening, upscaling and audio repair as not needing disclosure.
| Step | Sume feature | YouTube page treatment |
|---|---|---|
| Write the script | Your text or any writing tool | Exempt: script generation |
| Create the presenter | POST /v1/avatar-1.0/generate | Assess: realistic synthetic person |
| Approve the first frame | Avatar video preview | Not a publishing step |
| Render the video | POST /v1/avatar-1.0/talking-video | Assess: realistic content |
| Add captions | captions field or Video captions | Exempt: caption generation |
| Upscale or sharpen | Any enhancement tool | Exempt: enhancement |
Where Sume helps you decide
The preview endpoint creates first-frame stills before the full render. A reviewer can look at the stills and decide on disclosure before any publishing step. Preview stills are tier-independent, so you can approve and then pick a different quality tier for the final render without a new preview. Captions are stored at preview create and applied at generate-video, which keeps the caption text under review as well.
- Review the stills, not only the script.
- Record who approved the first frame.
- Write the disclosure into the script or first caption cue so it travels with the file.
What to do with the label itself
If your review says the video needs the label, you set it in YouTube Studio at upload, not in the Sume API. The YouTube page says that choosing "yes" adds a label to the video description, and that a creator who chooses no when the content needs a label may have one applied that cannot be removed. Keep the file's own on-screen or spoken disclosure as well, since it stays with the video if it is reposted.
A simple checklist
Before upload: confirm the presenter is not a real person who did not say the script, confirm any scene photo is not a deceptively altered real place, confirm the label setting, and keep the approval record. None of this is legal advice; it is a way to line your own process up with the platform page.
If you work on a team, make the checklist a gate rather than a reminder. The preview step is a natural place: nothing is rendered in full until a named person looks at the stills and marks the disclosure decision. Because changing the final render tier does not need a new preview, the reviewer's approval stays valid when finance asks you to move from max to plus.
Sources
Related posts
More in Sume Avatar 1.0
- Does a scripted AI presenter video need YouTube's synthetic label?
YouTube asks for a label when content makes a real person appear to say something they did not. Here is how to read that for a made-up Sume Avatar presenter.
- Will a YouTube AI label hurt reach? What the page says
YouTube's help page says disclosing AI content does not limit reach or monetization eligibility; penalties target non-disclosure. Plan an avatar series on that.
- Introducing Sume Avatar 1.0
Sume Avatar 1.0 is a multi-agent orchestration system as a single avatar model.
- Avatar Face Swap API (Beta): apply an avatar face to a video
Avatar Face Swap 1.0 is a Beta Sume endpoint that applies a ready avatar's face to a short public source video. Required fields, limits, and polling.
Written by Sume